Apparatus for generating temperature prediction model and method for providing simulation environment

Abstract

An apparatus for generating a temperature prediction model is disclosed. The apparatus for generating a temperature prediction model includes the temperature prediction model configured to provide a simulation environment, and a processor configured to set a hyperparameter of the temperature prediction model, train the temperature prediction model, in which the hyperparameter is set, so that the temperature prediction model, in which the hyperparameter is set, outputs a predicted temperature, update the hyperparameter on the basis of a difference between the predicted temperature, which is outputted from the trained temperature prediction model, and an actual temperature, and repeat the setting of the hyperparameter, the training of the temperature prediction model, and the updating of the hyperparameter on the basis of the difference between the predicted temperature and the actual temperature by a predetermined number of times or more to set a final hyperparameter of the temperature prediction model.

The invention claimed is: 
     
       1. An apparatus for generating a temperature prediction model, the apparatus comprising:
 the temperature prediction model configured to provide a simulation environment; and 
 a processor configured to:
 set a hyperparameter of the temperature prediction model; 
 train the temperature prediction model, in which the hyperparameter is set, so that the temperature prediction model, in which the hyperparameter is set, outputs a predicted temperature; 
 update the hyperparameter on the basis of a difference between the predicted temperature outputted from the trained temperature prediction model, and an actual temperature; and 
 repeat the setting of the hyperparameter, the training of the temperature prediction model, and the updating of the hyperparameter on the basis of the difference between the predicted temperature and the actual temperature by a predetermined number of times or more to set a final hyperparameter of the temperature prediction model, 
 
 wherein the hyperparameter is updated based on a first condition and a second condition, 
 the first condition is to update only a first element of the hyperparameter excluding a second element of the hyperparameter set to a fixed value by a user, and 
 the second condition is to maintain the fixed value for the second element of the hyperparameter without updating the second element of the hyperparameter, and 
 wherein the first element of the hyperparameter includes a number of layers and a number of times of repetition learning, 
 the second element of the hyperparameter includes a number of nodes for each layer, a learning rate, and a drop rate, and 
 the first element of the hyperparameter is searched by the processor when the second element of the hyperparameter is set to the fixed value by the user. 
 
     
     
       2. The apparatus according to  claim 1 , wherein the temperature prediction model is a recurrent neural network that is trained by using time series data comprising control information and a temperature according to the control information so as to output the predicted temperature. 
     
     
       3. The apparatus according to  claim 2 , wherein the processor is configured to:
 provide the time series data to the temperature prediction model in which the hyperparameter is set to train the temperature prediction model so that the temperature prediction model in which the hyperparameter is set outputs the predicted temperature; 
 input actual control information for a predetermined time period into the trained temperature prediction model; and 
 update the hyperparameter on the basis of a difference between the actual temperature corresponding to the actual control information for the predetermined time period and the predicted temperature outputted based on the actual control information for the predetermined time period. 
 
     
     
       4. The apparatus according to  claim 1 , wherein the processor is configured to set a hyperparameter, in which the difference between the predicted temperature outputted from the trained temperature prediction model and the actual temperature is minimized, to the final hyperparameter, within a searching range of the hyperparameter. 
     
     
       5. The apparatus according to  claim 1 , wherein the processor is configured to set a hyperparameter, in which the difference between the predicted temperature outputted from the trained temperature prediction model and the actual temperature is less than a preset value, to the final hyperparameter. 
     
     
       6. The apparatus according to  claim 1 , wherein the processor is configured to update the hyperparameter on the basis of any one algorithm of bayesian optimization, reinforcement learning, and bayesian optimization & hyperband. 
     
     
       7. The apparatus according to  claim 1 , wherein the temperature prediction model is configured to provide a simulation environment to an artificial intelligence device, and
 wherein the artificial intelligent device includes a neural network. 
 
     
     
       8. The apparatus according to  claim 7 , wherein the simulation environment includes outputting a value according to a degree of opening or closing of a valve. 
     
     
       9. A method for providing a simulation environment, the method comprising:
 setting a hyperparameter of a temperature prediction model, training the temperature prediction model, in which the hyperparameter is set, so that the temperature prediction model, in which the hyperparameter is set, outputs a predicted temperature, and updating the hyperparameter on the basis of a difference between the predicted temperature outputted from the trained temperature prediction model, and an actual temperature; and 
 repeating the setting of the hyperparameter, the training of the temperature prediction model, and the updating of the hyperparameter on the basis of the difference between the predicted temperature and the actual temperature by a predetermined number of times or more to set a final hyperparameter of the temperature prediction model, 
 wherein the hyperparameter is updated based on a first condition and a second condition, 
 the first condition is to update only a first element of the hyperparameter excluding a second element of the hyperparameter set to a fixed value by a user, and 
 the second condition is to maintain the fixed value for the second element of the hyperparameter without updating the second element of the hyperparameter, and 
 wherein the first element of the hyperparameter includes a number of layers and a number of times of repetition learning, 
 the second element of the hyperparameter includes a number of nodes for each layer, a learning rate, and a drop rate, and 
 the first element of the hyperparameter is searched by the processor when the second element of the hyperparameter is set to the fixed value by the user. 
 
     
     
       10. The method according to  claim 9 , wherein the temperature prediction model is a recurrent neural network that is trained by using time series data comprising control information and a temperature according to the control information so as to output the predicted temperature. 
     
     
       11. The method according to  claim 10 , wherein the updating of the hyperparameter comprises:
 providing the time series data to the temperature prediction model in which the hyperparameter is set to train the temperature prediction model so that the temperature prediction model in which the hyperparameter is set outputs the predicted temperature; 
 inputting actual control information for a predetermined time period into the trained temperature prediction model; and 
 updating the hyperparameter on the basis of a difference between the actual temperature corresponding to the actual control information for the predetermined time period and the predicted temperature outputted based on the actual control information for the predetermined time period. 
 
     
     
       12. The method according to  claim 9 , wherein the setting of the final hyperparameter of the temperature predication model comprises setting a hyperparameter, in which the difference between the predicted temperature outputted from the trained temperature prediction model and the actual temperature is minimized, as the final hyperparameter within a searching range of the hyperparameter. 
     
     
       13. The method according to  claim 9 , wherein the setting of the final hyperparameter of the temperature predication model comprises setting a hyperparameter, in which the difference between the predicted temperature outputted from the trained temperature prediction model and the actual temperature is less than a preset value, to the final hyperparameter. 
     
     
       14. The method according to  claim 9 , further comprising:
 inputting control information into the temperature prediction model, in which the final hyperparameter is set, to acquire the predicted temperature; and 
 allowing an artificial intelligence device to update, based on reinforcement learning, a control function on the basis of the predicted temperature corresponding to the control information. 
 
     
     
       15. The method according to  claim 9 , wherein the temperature prediction model provides a simulation environment to an artificial intelligence device, and wherein the artificial intelligent device includes a neural network. 
     
     
       16. The method according to  claim 15 , wherein the simulation environment includes outputting a value according to a degree of opening or closing of a valve.